Elasticsearch查询异常：搜索Shirt却返回含Short的商品结果

问题描述 #

搜索"Shirt"时，首位返回结果是名称包含"Short"的商品，以下是查询请求和响应详情。

查询请求 #

{"explain": true, "size": 3, "query": {"bool": {"filter": {"bool": {"must": [{"terms": {"visibility": [3,4]}}, {"terms": {"status": [0,1]}}, {"terms": {"stock.is_in_stock": [true]}}]}}, "must": {"function_score": {"functions": [{"filter": {"match": {"attribute_code": "attribute_value"}}, "weight": 1}], "score_mode": "multiply", "boost_mode": "multiply", "max_boost": 100, "min_score": 1, "query": {"bool": {"should": [{"multi_match": {"fields": ["name^2", "sku^50", "category.name^1"], "query": "shirt", "operator": "or", "fuzziness": 2, "cutoff_frequency": 0.01, "max_expansions": 3, "prefix_length": 2, "minimum_should_match": "75%", "tie_breaker": "1"}}, {"bool": {"should": [{"terms": {"configurable_children.sku": ["shirt"]}}, {"match_phrase": {"sku": {"query": "shirt", "boost": 1}}}, {"match_phrase": {"configurable_children.sku": {"query": "shirt", "boost": 1}}}]}}]}}}}}}

部分响应 #

{"took": 6, "timed_out": false, "_shards": {"total": 1, "successful": 1, "skipped": 0, "failed": 0}, "hits": {"total": {"value": 3, "relation": "eq"}, "max_score": 10.345283, "hits": [{"_shard": "[vue_storefront_catalog_1_product_1666220454][0]", "_node": "fM1XenujQqO0VZKK3bAdVw", "_index": "vue_storefront_catalog_1_product_1666220454", "_type": "product", "_id": "37522", "_score": 10.345283, "_source": {"...", "name": "MISAKI PENDANT W IDYLL SHORT", "image": "/7/4/74202_2_1_1_1.jpg", "small_image": "/7/4/74202_2_1_1_1.jpg", "thumbnail": "/7/4/74202_2_1_1_1.jpg", "msrp_display_actual_price_type": "0", "url_key": "misaki-pendant-w-idyll-short-37522", "...", "attributes_metadata": []}, "_explanation": {"value": 10.345283, "description": "sum of:", "details": [{"value": 10.345283, "description": "function score, product of:", "details": [{"value": 10.345283, "description": "sum of:", "details": [{"value": 10.345283, "description": "sum of:", "details": [{"value": 10.345283, "description": "weight(name:short in 16444) [PerFieldSimilarity], result of:", "details": [{"value": 10.345283, "description": "score(freq=1.0), computed as boost * idf * tf from:", "details": [{"value": 3.5200002, "description": "boost", "details": []}, {"value": 6.939326, "description": "idf, computed as log(1 + (N - n + 0.5) / (n + 0.5)) from:", "details": [{"value": 13, "description": "n, number of documents containing term", "details": []}, {"value": 13932, "description": "N, total number of documents with field", "details": []}]}, {"value": 0.42352825, "description": "tf, computed as freq / (freq + k1 * (1 - b + b * dl / avgdl)) from:", "details": [{"value": 1.0, "description": "freq, occurrences of term within document", "details": []}, {"value": 1.2, "description": "k1, term saturation parameter", "details": []}, {"value": 0.75, "description": "b, length normalization parameter", "details": []}, {"value": 5.0, "description": "dl, length of field", "details": []}, {"value": 4.2408123, "description": "avgdl, average length of field", "details": []}]}]}]}]}]}, {"value": 1.0, "description": "min of:", "details": [{"value": 1.0, "description": "No function matched", "details": []}, {"value": 100.0, "description": "maxBoost", "details": []}]}]}, {"value": 0.0, "description": "match on required clause, product of:", "details": [{"value": 0.0, "description": "# clause", "details": []}, {"value": 1.0, "description": "#visibility:{3 4} #status:{0 1} #ConstantScore(stock.is_in_stock:T)", "details": []}]}]}}]}}}

原因分析 #

从响应的_explanation字段能明确看到，这条结果的得分来自name:short的匹配。核心原因是查询中multi_match设置了fuzziness: 2——该参数允许查询词与文档词之间最多存在2个字符的编辑距离。"shirt"和"short"仅第2个字符不同（h vs o），编辑距离为1，完全符合模糊匹配的条件，因此被成功匹配。

另外，查询中的prefix_length:2要求前2个字符完全匹配，而"shirt"和"short"的前2个字符都是"sh"，满足该条件；minimum_should_match:75%针对单词查询时自动生效，所以这条包含"short"的文档被召回，且在没有其他更高得分文档的情况下，排在了结果首位。

内容的提问来源于stack exchange，提问作者SolarNewb